Home

ABSTRACT: In recent years, experimental correspondence studies have been frequently used to study labor market discrimination, but these only provide evidence on interview call-backs, leaving us without good evidence on important outcomes like wages, employment, and retirement. Such outcomes can be studied when discrimination laws change, but for most forms of discrimination there have not been changes since the 1960s, so whether these outcomes are impacted in the modern labor market is unknown. However, there is one form of discrimination where the law has changed in recent years: age discrimination. In 2009, an unexpected Supreme Court decision weakened federal age discrimination protections for older workers. In the fifteen years since, however, courts have ruled that protections are stronger under some states’ laws, thereby changing the age discrimination laws in those jurisdictions. We document these law changes via an extensive review of court cases in every state, and estimate whether their stronger protections improve employment outcomes for older workers. Using a synthetic difference-in-differences approach and data from the Current Population Survey, we find that the earnings of workers ages 50 to 75 increase by over a thousand dollars per year in response to the law changes. These effects are driven most strongly by middle-aged men, though women on the older end of our age range also experience increased earnings. We also find evidence of greater labor force attachment for both men and women, a result driven almost entirely by workers over age 65. Our results are consistent with age discrimination still affecting some older workers in the contemporary economy but show that stronger policies can provide protection and improve these workers’ employment outcomes.

Interpretation of Nonlinear Difference-in-differences: The Role of the Parallel Trends Assumption

June 2021 Version: SSRN, EconPapers

ABSTRACT: I argue interpretation of nonlinear difference-in-differences models depends on the form of the parallel trends assumption. When they are assumed in the natural scale of the dependent variable, the treatment effect is the interaction effect (a cross-difference). If they are assumed in the transformed scale, it is a single difference. I further note that assuming parallel trends in one scale implies they do not hold in the other, except in special cases. Finally, I consider log-linear (and related) difference-in-differences models and provide a constant form of the treatment effect that is comparable across applications with different parallel trends assumptions.

Work in Progress

Educational Impacts of Making College Tuition Free

With Joanne Song McLaughlin and Sirui Liu

Reduced Bias Estimation in Subjective Well-being Valuation of Health Conditions

Published & Accepted Articles

Does Knowing the Costs of Other Physicians Affect Doctors' Referrals?

Journal of Health Economics, August 2025.
Finalist for 2026 NIHCM Foundation Policy Research Award
Ungated accepted version
July 2021 Version: SSRN, EconPapers
Working paper circulated as, "Physician Response to Prices of Other Physicians: Evidence from a Field Experiment"

ABSTRACT: Patient referrals from primary care physicians (PCPs) to specialists are common in the American health care industry, but are typically made without any knowledge of relative specialist costs. In this study, I estimate the effect of providing such information to PCPs on referral patterns. Implementing a field experiment with an Independent Practice Association (IPA), I sent a list of average costs for new ophthalmology referrals to randomly chosen primary care medical practices. Using administrative referral data, I find that PCPs increased referral share to less costly ophthalmology practices during the first two months after treatment by 4.6 percentage points for each reduction in costliness rank (e.g., each rank closer to the least expensive). Effects were only found for patients for whom the PCPs had cost reduction incentives, and dissipated over the following four months. For the patients whose referrals were affected, I estimate that the expected cost to the IPA of a referral to ophthalmology fell during the first two months by about $80 (45% of pre-intervention referral cost).

Young Children and Parents' Labor Supply during COVID-19

American Journal of Health Economics, August 2025.
With Joanne Song McLaughlin and Yinlin Dai
Pre-registrations: main analysis; sub-analysis
Older (May 2021) Version: SSRN, EconPapers

ABSTRACT: We study the relationship between childcare needs during the COVID-19 pandemic and parental labor supply. Using monthly Current Population Survey data and following a pre-analysis plan, we implement three variations of an event-study research design comparing workers with varying levels of childcare responsibilities. The first compares parents with young children to adults without young children, while the second and third rely on the presence of someone who could provide childcare in the household: a teenager in one and a grandparent in the other. Across these approaches, we find childcare needs were not negatively associated with parents’ labor supply during the pandemic. We also do not find any difference in estimates between men and women. At the onset of the pandemic, many employers adopted flexible working arrangements. We provide evidence suggesting the ability to work remotely may have helped many parents avoid labor supply decreases.

Spillover Effects of Medicaid Expansion on Medicare: Evidence from Administrative Data

Journal of Policy Analysis and Management, March 2025.
With Dajung Jun and Yuting Zhang

ABSTRACT: The 2014 Medicaid expansion excluded Americans who were 65 years old and older, but they could still be affected via spillover effects. Using Medicare administrative data, we test for spillovers in Medicare spending and Medicaid coverage among low-income Medicare beneficiaries. We analyze two cohorts: those under 65 in 2014, who could have been induced by the expansion to take up Medicaid before joining Medicare; and those 65 or older in 2014, whose Medicaid eligibility was never affected by the expansion. We only find spillovers for the under-65 cohort, where Medicare spending fell and Medicaid coverage increased, with no measurable adverse effect on mortality. Combined with a null effect for the over-65 cohort, these facts suggest Medicare beneficiaries were not crowded out of health care by the expansion. Instead, those under-65 cohort satisfied “pent-up” demand via Medicaid, consuming care they would have otherwise obtained later under Medicare.

In Sickness and in Health: Interaction Effects of State and Federal Health Insurance Coverage Mandates on Marriage of Young Adults

Journal of Human Resources, March 2022.
With Joanne Song McLaughlin
Online appendix; Replication kit

ABSTRACT: We study the interaction of state and federal dependent health insurance mandates on young adult marriages. Using a new dataset on state-level mandates, we show marriage restrictions of these laws reduced marriage likelihoods by about two percentage points. When the Affordable Care Act (ACA) was enacted, its mandate ended marriage restrictions, encouraging marriage among those previously eligible for state mandates. However, among those ineligible for state mandates, it discouraged marrying to obtain insurance through spouses. The combination of these contrasting ACA effects eliminated the marriage gap. We also find these marriage effects resulted in corresponding impacts on out-of-wedlock births.

A Reevaluation of the Effects of State and ACA Dependent Coverage Mandates on Health Insurance Coverage

Journal of Policy Analysis and Management, Summer 2020.
With Joanne Song McLaughlin and Alex Ray
Replication kit
Visual summary: "JPAM at a Glance"

ABSTRACT: We study state and federal health insurance coverage mandates for young adults. Despite consistent findings that the Affordable Care Act's (ACA) federal mandate was effective, research has disagreed on whether pre-existing, state-level mandates were successful in increasing coverage. We reconsider the issue with a new analytical perspective and newly available, accurate data on state mandates. We show that the impact of the state mandates was substantive and concentrated among young adults between ages 19 and 23. Our estimates indicate that the dependent coverage rose by 3.9 percentage points and overall coverage rose by 3.3 percentage points. Crowd out of coverage through young adults' own jobs was negligible. For those above 23, we find little evidence of changes in coverage. We incorporate these insights into analysis of the ACA's mandate, showing its effects were focused among those who were not eligible for state mandates, or were eligible but older than 23. Our results suggest eligibility restrictions played important roles in limiting the scope of the state mandates, but they can be practical and effective tools for policymakers looking to ensure or expand coverage for young adults in the face of uncertainty about the ACA.

Does Government Health Insurance Reduce Job Lock and Job Push?

Southern Economic Journal, July 2020.

ABSTRACT: I study job lock and job push, twin phenomena believed to be partially due to employment-contingent health insurance (ECHI). Using variation in Medicaid eligibility among household members of male workers to identify changes in those workers’ reliance on ECHI, I estimate notable job lock and job push effects. For married male workers, a 15 percentage point increase in the likelihood a household member is eligible for Medicaid increases the rate of voluntary job exits over a four-month period by 14 percent. For job push, the same increase in a household member’s likelihood of Medicaid eligibility reduces the transition rate into jobs with ECHI among all male workers by 8 percent.

The Effect of Health Insurance on Crime: Evidence from the Affordable Care Act Medicaid Expansion

Health Economics, March 2020.
With Qiwei He

ABSTRACT: Little evidence exists on the effect of the Affordable Care Act (ACA) on criminal behavior, a gap in the literature that this paper seeks to address. Using a simple model, we argue we should anticipate a decrease in time devoted to criminal activities in response to the expansion, since the availability of the ACA Medicaid coverage raises the opportunity cost of crime. This prediction is particularly relevant for the ACA expansion since it primarily affects childless adults, a population likely to contain individuals who engage in criminal behavior. We validate this forecast empirically using a difference-in-differences framework, estimating the expansion's effects on panel datasets of state- and county-level crime rates. Our estimates suggest that the ACA Medicaid expansion was negatively associated with burglary, vehicle theft, homicide, robbery, and assault. These crime-reduction spillover effects represent an important offset to the government's cost burden for the ACA Medicaid expansion.

Does Regulation of Physicians Reduce Health Care Spending?

Southern Economic Journal, April 2017.
Winner of the 2017 Georgescu-Roegen Prize: awarded annually by the Southern Economic Association to the best academic article in the Southern Economic Journal.
Replication kit

ABSTRACT: The medical community argues that physician fear of legal liability increases health care spending. Theoretically, though, the effect could be positive or negative, and empirical evidence has supported both cases. Previous studies, however, have ignored the fact that physicians face risk from industry oversight groups like state-level medical licensing boards in addition to civil litigation risk. This paper addresses this omission by incorporating previously unused data on punishments by oversight groups against physicians, known as adverse actions, along with malpractice payments data to study state-level health care spending. My analysis suggests that, contrary to conventional wisdom, spending does not rise in response to increased risk. An increase in adverse actions of 16 (the year-to-year average) is associated with statistically significant, annual decreases in state spending on hospital care of approximately $22 million, and on prescription drugs of nearly $10 million. Malpractice payments are estimated to have smaller, statistically insignificant effects.

Does Employer-provided Health Insurance Constrain Labor Supply Adjustments to Health Shocks? New Evidence on Women Diagnosed with Breast Cancer

Journal of Health Economics, September 2013.
With Cathy J. Bradley and David Neumark
Media coverage: Slate.com's Moneybox blog.

ABSTRACT: Employment-contingent health insurance may create incentives for ill workers to remain employed at a sufficient level (usually full-time) to maintain access to health insurance coverage. We study employed married women, comparing the labor supply responses to new breast cancer diagnoses of women dependent on their own employment for health insurance with the responses of women who are less dependent on their own employment for health insurance, because of actual or potential access to health insurance through their spouse’s employer. We find evidence that women who depend on their own job for health insurance reduce their labor supply by less after a diagnosis of breast cancer. In the estimates that best control for unobservables associated with health insurance status, the hours reduction for women who continue to work is 8 to 11 percent smaller. Women’s subjective responses to questions about working more to maintain health insurance are consistent with the conclusions from observed behavior.

Non-Academic Writing

The NBA Should Use The Salary Cap To Combat Tanking

Page updated

Google Sites

Report abuse